Using a transliteration to spellcheck a syllabic script?

Question

I’m working in a context where we are standardizing spellings, and I’m just exploring some of the spellchecking tools in PT8 before assuming we have to do it manually. Currently length and pre-aspiration are variably specified, leading to a lot of spelling variants.

Algonquian languages are difficult to spellcheck because they have thousands of inflectional endings. I would be able to use the PT8 morphological spellchecker to isolate roots, except we use a syllabic writing system that obscures the morphemes boundaries between a consonant and a vowel.

Has anyone successfully used any Paratext or FLEx tools in such a situation? I easily created a transliterated roman project, and I could use the morphological spellchecker there. But I don’t know how it would work out applying changes back to the main project in syllabics.

Other Algonquian languages have worked in Roman orthography and then converted afterward to the syllabic scripts, but I work in a region where our translators are currently literate in the syllabic script only.

Any ideas would be welcome,

Answer 1

MW,

The only other situation like this I have seen is languages using the Amharic script which is also a syllabary. Words could be checked in Flex with either the morphology parser defining “stem names”, or using the phonological parser “hermit crab parser”. I seem to remember that phonological parser was better, but has a steep learning curve. Maybe some one working in Ethiopia will share with you what they have learned.

Answer 2

About 10 years ago, my role involved training and support for Paratext users, but I have not used it much since then, so I can’t contribute much to the Paratext aspect of your question. My current role involves helping FLEx users implement one or the other of the automatic parsers to their project, so I’m responding to this post from this perspective.

Spellchecking an agglutinative language such as Algonquian is indeed a challenge, given the tremendously large number of affix combinations possible in languages like that. The language I worked in in Africa does not compare in that regard, but other complexities made it such that the team using Paratext never found a reliable way to do spellchecking in Paratext. I configured one of the FLEx parsers to decode morphophonemic and tonal changes and used it to spellcheck the biblical text prior to publication of the NT, but corrections had to be made manually in Paratext. We did not find a way to harness the power of the FLEx parser automatically in Paratext.

If doing something like that is of interest to you for Algonquian, I am available to help with the configuration of the FLEx parser. In order for it to work, the parser would need either (1) to have an alphabetic (non-syllabic) version of the orthography to process or (2) to itself produce an alphabetic form to work with and then convert the result back into the syllabic orthography.

Contact me at [Email Removed] if you want more information or are interested in my help.

Answer 3

Ok, thanks anon044949 & anon962715! I’m pretty sure an Ojibwe specialist I know created a FLEx parser (and some other kind of third party software parser) for the language, which would be very similar to the one I would need. I’ll look into it.